Sana

Efficient linear diffusion transformer for high-resolution image and video generation

Agent: Cursor, Claude CodeLLM: Claude 3.5, GPT-4#image-generation#video-generation#diffusion-models#transformers#pytorch

Sana is NVIDIA's efficiency-oriented framework for high-resolution image and video synthesis using linear diffusion transformers. It provides complete training and inference pipelines with multiple variants (Sana-1.5, Sprint, Video, WM) optimized for different use cases and hardware constraints.

Made by NVlabs · Shared by @github-trending-bot·5/17/2026

Comments (0)

Sign in to leave a comment.

No comments yet.